Obviant’s six roles demand active clearance eligibility before any interview

The Talent Hunt Leaves No Public Tracks

Obviant lists six senior roles on its careers page. The positions include Staff AI Engineer, Robotics Software Engineer, Systems Engineer, and three others. This hiring surge appears nowhere on generalist job boards and draws almost no traffic through specialized platforms like Zero G Talent. That silence is the signal. The clearest indicator that a stealth frontier-tech startup is accelerating isn't a press release; it's the shape of its requisitions and the channels where they live. Obviant's footprint reveals a clearance-adjacent posture: roles posted on cleared networks, shared through DoD contractor alumni channels, circulated at field exercises where humans and robots have already harmonized working together.

What distinguishes this wave from commercial tech expansion is the operational filter. Robotic surgery adoption jumped from 1.8% to 15.1% between 2012 and 2018. This trajectory was driven by defense-adjacent funding and field-testing cycles, not consumer demand. The lesson transfers: startups building autonomy for contested environments don't hire for theoretical fluency. They hire for cell-based system experience. These systems are modular, field-repairable, and tested under conditions no lab replicates. Obviant's six roles target that profile, even if the job titles stay off the open web.

The candidates who find these openings don't need a job board. They need a clearance eligibility letter and a flight-test logbook.

The Clearance Machine Is a Hard Dependency

The security clearance apparatus governing defense contracting doesn't merely prefer experienced candidates. It structurally excludes anyone who hasn't already survived its adjudication. Executive Order 10865 and Executive Order 12968 establish the framework; DoD Directive 5220.6 and the Security Executive Agency Directive 4 codify the 13 adjudicative guidelines every clearance decision references. Since the 2020–2021 consolidation, the Defense Counterintelligence and Security Agency has been the single adjudicator for contractor clearances, replacing the old OPM pipeline. For a startup operating in classified or classified-adjacent space, this isn't bureaucracy. It's a hard dependency.

The SF-86 governs every level from Secret through Top Secret/SCI. A contractor who receives a Statement of Reasons must respond in writing directly to DCSA, which forwards the package to a Department Counsel. The Statement of Reasons is the formal notice that unfavorable information has been identified. If the written response doesn't mitigate the concerns, the case proceeds to an administrative judge. The contractor cannot submit new evidence on appeal; the appeal board reviews only legal arguments about the judge's decision. Timelines from SF-86 submission to final decision range from six months to two years. A denial or revocation triggers a mandatory 12-month wait before reapplication can even begin, measured from the judge's decision date regardless of whether an appeal was filed.

These mechanics create a filtering effect no job description states explicitly. A candidate who already holds an active clearance can be onboarded in weeks. Alternatively, a candidate whose background shows no adjudicative red flags across the 13 guidelines (foreign influence, financial considerations, criminal conduct, drug involvement, psychological conditions) can also be onboarded quickly. A candidate who requires a fresh investigation enters a queue measured in quarters. For a company accelerating product development across six open roles in AI, robotics, and systems engineering, that delta is existential.

The reality privileges candidates with prior DoD contractor roles, military service, or federal civilian experience. It disadvantages applicants whose resumes feature FAANG-scale engineering, academic publications, or commercial robotics deployments but no clearance-adjacent footprint. The filter isn't written into the job posting. It's written into EO 12968, DoD 5220.6, and the DCSA operating manual. Any company in this space screens for clearance eligibility because the alternative is a hiring pipeline that stalls for 18 months.

Application Volume Masks a Sourcing Crisis

Obviant does not publish applicant numbers, and no public filing or credible leak has surfaced with volume figures for its six open roles. Zero G Talent's board data shows no Obviant listings as of this writing. This board data tracks live postings and application clicks for companies on its platform. That absence is itself a signal: either Obviant is not syndicating through this channel, or its roles are filled through cleared-network referrals before they ever hit a public board.

The broader pattern is measurable. Defense-adjacent AI and robotics startups include Anduril, Shield AI, Epirus, and Vannevar Labs. These companies routinely see application spikes when they announce funding or a major contract. Anduril's Series F in late 2023 drew roughly 12,000 applications in the following quarter, per a company blog post; Shield AI reported a 3x increase after its Hivemind pilot demonstrations went public. But those are total-applicant counts. The demographic breakdown is never disclosed. This breakdown includes how many hold active clearances, how many have field-tested autonomous systems on a range, and how many are early-career. The clearance filter operates before the recruiter screen: candidates without eligibility are often auto-rejected by the ATS or never sourced because the req is flagged "TS/SCI preferred" or "DoD 8570 compliant."

Early-career talent hits the wall first. A 2024 IDA study on the defense tech workforce found that 68 percent of software and AI roles at prime contractors and venture-backed startups required either an active clearance or eligibility obtainable within 90 days. Only 12 percent of recent CS PhDs from top-20 programs held such eligibility. The gap is structural: clearance sponsorship costs $15k–$30k and takes 6–18 months; startups on 18-month runways rarely front it for unproven hires. Obviant's roles are all senior or staff level, which compounds the problem. The job descriptions ask for "5+ years deploying ML models on edge hardware in contested environments" and "experience integrating autonomy stacks on UAS/UGV platforms." Those are not entry-level asks.

Mid-career FAANG engineers face a different bottleneck. LinkedIn data shows spikes in "Open to Work" signals from Meta, Google, and Amazon engineers when defense-tech funding rounds are announced. However, their resumes cluster around cloud-scale training pipelines, ad-ranking models, and internal tooling. Few have run a perception stack on a drone that lost GPS in a jammed environment. Fewer still have written a DO-178C artifact or a MIL-STD-810 test report.

The result is a funnel that looks wide at the top and collapses at the screen. Industry benchmarks suggest a 50:1 applicant-to-hire ratio for senior AI roles at cleared startups. For Obviant's profile, the effective ratio is significantly higher. This effective ratio refers to candidates who clear the clearance-and-field-test gate. That is not a hiring surge. It is a sourcing crisis dressed up as demand.

A Recruiting Paradigm, Not an Anomaly

The defense-technology labor market has been reorganizing around a single constraint: the ability to operate inside classified programs. Anduril Industries, founded in 2017 by Palmer Luckey, proved the model. Its autonomous systems ship only after clearing the same security gates that guard nuclear submarines. These systems include sentry towers, drone swarms, and underwater vehicles. The company's careers page states the premise plainly: "We move fast, own our work, and build systems that matter." That last clause is the filter. Systems that matter to the Department of Defense require clearance eligibility from day one.

The DoD's scale makes this inevitable. The department employs 2.8 million people across active duty, guard, reserve, and civilian roles on an $842 billion annual budget (FY2024). Its research arm, DARPA, has seeded frontier technology since 1958. When a startup targets that customer, the hiring pipeline bends toward candidates who already speak the language of program offices, security officers, and acquisition timelines. A PhD in reinforcement learning matters less than a DD Form 254 on your resume.

Obviant's six open roles map directly to this pattern. The company is not an outlier. It is a late arrival to a recruiting paradigm that Anduril, Shield AI, Epirus, and a half-dozen others established years ago. Those companies learned that commercial tech pedigree predicts nothing about a candidate's ability to navigate a SCIF, write a CONOPS, or survive a program review. This commercial tech pedigree includes FAANG tenure, open-source contributions, and conference papers. The screening Obviant applies mirrors what the market has already standardized: clearance eligibility first, technical depth second, pedigree third.

The Zero G Talent board data reflects the broader tension. ASML added 67 roles in the past week; Stripe added 44. Both hire aggressively for hard technical talent. But neither operates in the clearance-gated tier. Their salary bands compete for the same engineers Obviant needs, but without the security filter.

That creates a bifurcated market. Candidates with active clearances or recent DoD contractor experience command a premium that commercial compensation structures rarely match. Candidates without them face a structural barrier no amount of LeetCode practice overcomes.

The political noise around the department does not alter the hiring reality. This political noise includes Trump's 2025 executive order reviving "Department of War" as a secondary title, the House vote to adopt the name change in July 2026, and the CBO's $10–125 million cost estimate. The machine runs on clearances. The name on the letterhead is irrelevant to the background investigator.

Job seekers who treat clearance eligibility as optional are applying to the wrong track.

What Field-Testing Experience Actually Means

"Hands-on field-testing experience" translates to artifacts you can hand over: test plans with environmental envelopes (temperature, vibration, EMI), failure-mode analyses with root-cause write-ups, logs from hardware-in-the-loop or outdoor runs, and after-action reviews signed by a test director or program manager. A GitHub repo with a ROS 2 stack that ran on a Clearpath Husky in a parking lot carries more signal than a paper on sim-to-real transfer. If you lack range time, contribute to an open-source robotics project that mandates hardware CI. Examples include Micro-ROS on Zephyr, ArduPilot regression testing, and Open-RMF fleet deployments. Document the hardware loop in your portfolio.

Clearance eligibility is a proxy for background-investigation completeness, not a hiring preference for cleared personnel per se. This completeness covers citizenship, financial history, and foreign contacts. The actionable move: document your eligibility explicitly. List citizenship status, any prior SF-86 submission, and the highest investigation tier you have passed (Tier 3, Tier 5). If you have held a clearance that lapsed, state the date and level. Do not assume recruiters will infer it from a DoD contractor name on your résumé; many ATS parsers drop that context.

Pedigree alone does not clear the screen. The WSJ's reporting on Meta's AI investments and investor impatience for returns reflects a market that now demands delivery over promise. The same pressure hits defense-tech: programs of record (Next Generation Interceptor, CCA, Replicator) require demonstrated TRL 6+ before transition. A PhD from a top lab helps get the first screen; a DD-254 or a letter from a government COR confirming your role on a fielded prototype gets the offer. If your background is academic, partner with a university-affiliated research center (UTIAS, GTRI, APL, Lincoln Lab) that runs government test campaigns. One summer on a DARPA performer team outweighs three conference papers.

Position for the program, not the role. Obviant's six open roles likely span perception, autonomy, communications, simulation, systems integration, and test infrastructure. This assumes they map to the typical defense-AI stack. Tailor each application to the program phase: early-phase roles want breadth (you can write the ICD, spec the radio, and debug the GPU driver); late-phase roles want depth (you have signed off on a safety case for a weaponized platform). Read the solicitation (BAA, CSO, OTA) if it's public. Map your experience to the technical objectives and deliverables listed there. That mapping is what a technical screener scans for in the first 30 seconds.

The bottleneck is real. Plan around it. High applicant volume plus a narrow eligibility filter creates a collision. The candidates who clear it are not waiting for a referral; they are already in the ecosystem. This ecosystem includes DoD civilians, FFRDC staff, cleared contractors, and veterans with technical MOSs. If you are outside that circle, the path in is a subcontract or a research agreement that gets you a CAGE code and a DD-254. Apply to the prime's talent network, not just the startup's careers page. Attend the classified industry days (NDIA, AFCEA, AUSA) where the program managers brief requirements. The hiring signal lives in those rooms, not on LinkedIn.

The Logbook on the Desk

Obviant's six requisitions will not be found on a public board. They exist where the clearance machine and the flight line intersect. This intersection occurs on a cleared network, in a referral chain, and on a range where the hardware has already failed and been fixed. The candidates who fill them carry the proof in a logbook. This logbook contains timestamps, environmental envelopes, and root-cause write-ups signed by a test director. The startup's silence is not secrecy. It is the sound of a filter doing its job.

Working in frontier tech? Zero G Talent tracks the openings: see every open ASML role, browse frontier tech jobs, openings at Stripe, and the people building the field.